The original report of the SNARC effect (Spatial Numerical Association Response Code) (Dehaene, Bossini, & Giraux, 1993 ) showed that in a parity judgment task, small numbers were responded to faster with the left hand while large numbers were responded to faster with the right hand. According to Dehaene et al. (1993), the SNARC effect depends on the quantitative representation of number, such as a left-to-right-oriented analog number line. Berch, Foley, Hill, and Ryan (1999) performed the first study to investigate SNARC effect in children by using a parity judgment task. SNARC effect was observed from 9 years on and it decreased with increasing age. As for the 7 years olds, there was no SNARC effect in number comparison task. Existing findings indicate that SNARC effect can be influenced by children's growing numerical skills and task demands. By using number comparison task and area comparison task that are relevant to magnitude，we set the main goal of current study as to exploring whether the children who could recognize Arabic numbers, but meanwhile have not familiar with numbers because of lacking in enormous amount of mathematical training, could demonstrate characteristics of spatial encoding in representation of numbers or area. We hypothesized that spatial representation could be automatically activated when 5 year-olds representing magnitude of circle area rather than magnitude of Arabic number. In order to decrease the influence of mathematical training, we selected the children in the middle class in kindergarten as subjects for they could discriminate Arabic numbers but not very familiar with them. There were two kinds of task. The first one was number comparison task which was adopted by van Galen & Reitsma（2008）who had found that 7 years olds could activate mental number line. We anticipated there would not SNARC effect in number comparison task for the children were far below 7 years old. The second task was area comparison task which was easier than number comparison task for the children were more familiar with area than numbers in their daily life. And we anticipated that in the second task, SNARC effect would appear. Rts data of number comparison task analysis indicated that there was no SNARC effect，but distance effect was found. The analysis of error rate indicated that there was no SNARC effect or distance effect. The analysis of Rts data of circle area comparison task indicated that there were SNARC effect and distance effect. Right hand responded faster to relatively larger circles than to relatively smaller circles, while left hand responded faster to relatively smaller circles than to relatively larger circles. Error rate analysis also indicated that there existed SNARC effect but no distance effect. We could infer from these findings that early before the spatial coding of Arabic number, children might have represented non-symbolic magnitude spatially.